In the first installment of this article, we looked at some straightforward approaches to improve decision making about marketing resource allocation. In this second part, I'll draw from the best practices in organizations of all sizes to expose some additional opportunities for marketers today.
1. Capture Fragmented Spend in New Marketing Mix Models
Marketing mix modeling once was the preserve of only consumer packaged goods (CPG) marketers with big advertising budgets. Today, our most sophisticated antenna, employed with the latest modeling techniques and huge amounts of highly fragmented data allow B-to-B marketers to benefit as well. One example: a global technology leader today is able to map—across the US and China—thousands of small B-to-B spends with thousands of small spends in specialist digital and print spots as an input, and thousands of customers and their transactions as an output.
2. Understand How Intermediation Is Evolving
The best models capture the intermediated effects of marketing spend and how different marketing levers work. This can be seen in different ways. For example, automotive advertising might be intended to drive website traffic and dealer footfall in the short term that will be seen in stronger sales further in time. One TelCo company looks at how marketing spend drives Google queries, footfall in their stores, calls to the call center, traffic to their website, and ultimately improved retention and accelerated customer acquisition. The role of social media as an accelerant is also captured in these models, and marketing campaigns' effects can be dramatically positively or negatively moderated by this intermediation layer of social media.
Another example of intermediation is in the U.S. insurance market, where the traditional agent-based business model has come under threat of Web-based aggregators or direct business models. The way in which intermediation is changing from agents acting as a layer between the advertiser and the consumer, and now a more direct relationship, has to be tracked and optimized. There are huge marketing optimization opportunities in this space of changing and evolving business models in financial services, and of course, in retail.
3. Meta Models Include Many Separate Models and Research
You'll now be faced with so much research, Web-focused analytics, traditional marketing mix models and so on. But, they'll all be telling you a part of the story in isolation. Meta models can include multiple vendors' models and research in a single business simulation tool. One example is a 10-year, $2 trillion simulation for a U.S. retail gasoline company. This employs strategic models and tools to analyze marketing decisions with a long-term focus, often using customer-level CLTV models to analyze the impact of marketing at the customer level. Agent-based modeling uses similar techniques in computer games to build populations of customers, each with attributes that match real consumer behavior helps us create meta models that look at the long-term outcomes of marketing spend. A marketing mix model is no longer a single regression equation. It's a flexible system that can include multiple inputs from all your various sources of research, insight, and analysis.